Authors: KD Conway, Cathie So, Xiaohang Yu, Kartin Wong

Published on: January 31, 2024

Impact Score: 8.38

Arxiv code: Arxiv:2401.17555

Summary

• What is new: Introduction of opML, integrating machine learning with blockchain, enabling AI model inference directly on-chain.
• Why this is important: The need for decentralized, secure, and transparent AI services in blockchain systems.
• What the research proposes: opML uses an interactive fraud proof protocol for decentralized and verifiable consensus, enabling execution of large AI models on standard PCs.
• Results: Highly efficient and cost-effective ML services, with the capability to run extensive language models like 7B-LLaMA on PCs without GPUs.

Technical Details

Technological frameworks used: Blockchain, Optimistic Rollup systems

Models used: 7B-LLaMA

Data used: nan

Potential Impact

Companies in AI services, blockchain technology, and cloud computing markets.

Want to implement this idea in a business?

We have generated a startup concept here: ChainMind.